Unverified50% confidenceFactExact time
The Chinchilla empirical ratio is about 20 tokens per parameter.
1
Sources
50%
Confidence
Long-term
Relevance
11/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedChinchilla定律大致按每个参数分配20倍token的比例来安排算力75% similarUnverifiedThe rule of thumb given by the Chinchilla scaling law is that the number of training tokens should be approximately 20 times the number of parameters72% similarVerifiedIn 2022, Hoffmann et al.'s Chinchilla paper redid experiments on four hundred models and concluded that, under the premise of compute-optimality, each parameter should correspond to approximately 20 tokens.72% similarUnverifiedEach rollout simultaneously serves as a Thompson sampling trial over approximately 20 inference parameters.70% similarUnverifiedThe Chinchilla study found that under a fixed compute budget, the number of parameters and the number of training tokens should be matched at a roughly 1:20 ratio.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1168678API
curl https://kongchang.com/api/v1/knowledge/claims/1168678MCP
get_claim(id=1168678)